地球信息科学学报 ›› 2022, Vol. 24 ›› Issue (7): 1286-1300.doi: 10.12082/dqxxkx.2022.210823

• 地理空间分析综合应用 • 上一篇    下一篇

基于随机森林模型的中国中老年人群HDL-C环境影响因素研究

裴泽华1(), 葛淼1,*(), 李浩1, 何进伟2, 王聪霞3   

  1. 1.陕西师范大学地理科学与旅游学院健康地理研究所,西安 710119
    2.延安大学医学院,延安 716000
    3.西安交通大学医学院第二附属医院心内科,西安 710004
  • 收稿日期:2021-12-22 修回日期:2022-03-22 出版日期:2022-07-25 发布日期:2022-09-25
  • 通讯作者: * 葛 淼(1960—),男,陕西咸阳人,教授,主要从事健康地理研究。E-mail: gemiao@snnu.edu.cn
  • 作者简介:裴泽华(1997—),男,山西长治人,硕士生,主要从事健康地理研究。E-mail: pzh15635445598@163.com
  • 基金资助:
    国家自然科学基金项目(41761100)

Environmental Factors Influencing HDL-C in Middle-aged and Elderly Chinese Population based on Random Forest Model

PEI Zehua1(), GE Miao1,*(), LI Hao1, HE Jinwei2, WANG Congxia3   

  1. 1. Institute of Health Geography, School of Geography and Tourism, Shaanxi Normal University, Xi'an 710119, China
    2. Medical College of Yan'an University,Yan'an 716000, China 3. The Second Affiliated Hospital of Xi'an Jiaotong University, Xi'an 710004, China
    3. The Second Affiliated Hospital of Xi'an Jiaotong University, Xi'an 710004, China
  • Received:2021-12-22 Revised:2022-03-22 Online:2022-07-25 Published:2022-09-25
  • Contact: GE Miao
  • Supported by:
    National Natural Science Foundation of China(41761100)

摘要:

高密度脂蛋白胆固醇(HDL-C)可以有效促进人体内胆固醇的代谢外排,其水平的高低与患心血管疾病的风险呈负相关关系,是心血管疾病的预防与保护因素。厘清我国中老年人群HDL-C水平的地理分异特征及环境影响因素,对我国心血管疾病防治工作的开展有重要意义。论文基于中国中老年人纵向追踪调查,利用全局空间自相关、冷热点分析等方法阐释中国中老年人群HDL-C水平的空间分异特征及变化趋势,同时对比引入随机森林回归模型及多元线性回归方法探讨HDL-C水平空间分布的环境影响因素及其指示作用。结果表明:中国中老年人群HDL-C水平表现为女性高于男性、农村高于城镇,具有明显的地域差异性,整体呈现出“北低南高,中间过渡”的分布格局,且北方出现了以内蒙古、河北、辽宁为代表的低值聚集区,南方出现了以广东、广西、云南为代表的高值聚集区;SO2、NO2、降水、气压、PM10和PM2.5是影响中老年人群HDL-C水平差异分布的主要环境因素,其中高浓度的空气污染物是造成HDL-C值较低的危险因素,充沛的降水和低压环境是防治HDL-C值较低的保护因素。因此,今后关于HDL-C血脂异常防控工作在全国各地应注重其空间分布规律,重点加强对HDL-C低值区的监测,以达到因地制宜、精准防控的目的。

关键词: 中老年人群, 高密度脂蛋白胆固醇, 时空变化特征, 气象条件, 空气污染, 空间相关分析, 随机森林, 影响因素

Abstract:

High Density Lipoprotein Cholesterol (HDL-C) can effectively promote the metabolic efflux of cholesterol in human body, and its level is negatively correlated with the risk of cardiovascular disease, which is a preventive and protective factor of cardiovascular disease. It is of great significance to clarify the geographical characteristics and environmental factors of HDL-C level in middle-aged and elderly population in China. Based on the longitudinal survey of middle-aged and elderly people in China, this paper uses global spatial autocorrelation and cold hot spot analysis to explain the spatial characteristics and trends of HDL-C levels in middle-aged and elderly people in China. At the same time, the random forest regression model and multiple linear regression method are compared to explore the environmental factors influencing the spatial distribution of HDL-C level. The results show that the HDL-C level of middle-aged and elderly population in China is higher in females than that in males, and higher in rural areas than that in urban areas, with obvious regional differences. The overall distribution pattern is "low in the north and high in the south, with transition in the middle". In addition, there are low value aggregation areas in Inner Mongolia, Hebei, and Liaoning in the north and high value aggregation areas in Guangdong, Guangxi, and Yunnan in the south. The SO2, NO2, precipitation, air pressure, PM10 and PM2.5 are the main environmental factors affecting the different distributions of HDL-C level in middle-aged and elderly population. Among them, high concentration of air pollutants is the risk factor of low HDL-C value, while abundant precipitation and low-pressure environment are the protective factors to prevent and control low HDL-C value. Therefore, the prevention and control of HDL-C dyslipidemia should pay attention to its spatial distribution throughout the country in the future, focusing on strengthening the monitoring of HDL-C low value areas, so as to achieve the purpose of adjusting measures to local conditions and accurate prevention and control.

Key words: middle-aged and old people, high-density lipoprotein cholesterol, spatial-temporal variation characteristics, meteorological conditions, air pollution, spatial correlation analysis, random forest, influence factor